Subgroup Identification Analysis

GBSG Example

Author

Larry Leon

Code

# Set options
knitr::opts_chunk$set(
  echo = TRUE,
  warning = FALSE,
  message = FALSE,
  fig.align = 'center',
  fig.retina = 2
)
rm(list=ls())
library(tinytex)

Warning: package 'tinytex' was built under R version 4.5.2

Code

library(ggplot2)

#library(table1)

library(gt)

library(survival)
library(data.table)
library(randomForest)
library(grf)
library(policytree)
library(DiagrammeR)

#library(grid)
#library(forestploter)
#library(randomizr)

# library(devtools)
# install_github("larry-leon/weightedsurv", force = TRUE)
#install.packages("weightedsurv")
# install_github("larry-leon/forestsearch", force = TRUE)

library(forestsearch)
library(weightedsurv)

# Set theme for plots
theme_set(theme_minimal(base_size = 12))

1 Summary

Reproducing main GBSG analysis

1.1 Datasetup

Code

df.analysis <- gbsg
df.analysis <- within(df.analysis,{
id <- as.numeric(c(1:nrow(df.analysis)))  
# time to months
time_months <- rfstime/30.4375
grade3 <- ifelse(grade=="3",1,0)
treat <- hormon
})
confounders.name <- c("age","meno","size","grade3","nodes","pgr","er")
outcome.name <- c("time_months")
event.name <- c("status")
id.name <- c("id")
treat.name <- c("hormon")

1.2 Kaplan-Meier curves and baseline summary

Code

dfcount <- df_counting(
  df = df.analysis,
  by.risk = 6,
  tte.name = outcome.name, 
  event.name = event.name, 
  treat.name = treat.name
)
plot_weighted_km(dfcount, conf.int = TRUE, show.logrank = TRUE, ymax = 1.05, xmed.fraction = 0.775, ymed.offset = 0.125)

Code

create_summary_table(data = df.analysis, treat_var = treat.name, 
                     table_title = "GBSG Characteristics by Treatment Arm",
                                      vars_continuous=c("age","nodes","size","er","pgr"),
                                      vars_categorical=c("grade","grade3"),
                                      font_size = 12)

GBSG Characteristics by Treatment Arm
Characteristic		Control (n=440)	Treatment (n=246)	P-value¹	SMD²
age	Mean (SD)	51.1 (10.0)	56.6 (9.4)	<0.001	0.57
nodes	Mean (SD)	4.9 (5.6)	5.1 (5.3)	0.665	0.03
size	Mean (SD)	29.6 (14.4)	28.8 (14.1)	0.470	0.06
er	Mean (SD)	79.7 (124.2)	125.8 (191.1)	<0.001	0.30
pgr	Mean (SD)	102.0 (170.0)	124.3 (249.7)	0.213	0.11
grade				0.273	0.06
	1	48 (10.9%)	33 (13.4%)
	2	281 (63.9%)	163 (66.3%)
	3	111 (25.2%)	50 (20.3%)
grade3				0.174	0.05
	0	329 (74.8%)	196 (79.7%)
	1	111 (25.2%)	50 (20.3%)
¹ P-values: t-test for continuous, chi-square/Fisher's exact for categorical/binary variables
² SMD = Standardized mean difference (Cohen's d for continuous, Cramer's V for categorical)

1.3 GRF analysis

Code

## GRF
grf_est1 <- grf.subg.harm.survival(data=df.analysis,
confounders.name = confounders.name,
outcome.name=outcome.name, event.name=event.name, id.name=id.name, treat.name=treat.name,
maxdepth = 2, n.min = 60, dmin.grf = 12, frac.tau=0.6, details=TRUE)

tau, maxdepth = 46.75811 2 
   leaf.node control.mean control.size control.se depth
1          2         6.49        82.00       3.34     1
2          3        -4.10       604.00       1.06     1
11         4        -7.90       112.00       2.81     2
21         5         3.86       177.00       1.87     2
4          7        -5.89       356.00       1.33     2

Selected subgroup:
  leaf.node control.mean control.size control.se depth
1         2         6.49        82.00       3.34     1

GRF subgroup found
Terminating node at max.diff (sg.harm.id):
[1] "er <= 0"

All splits:
[1] "er <= 0"   "age <= 50" "age <= 43"

Code

# NOTE: In general for GRF trees
# leaf1 --> recommend control
# leaf2 --> recommend treatment
# Tree depth 1
plot(grf_est1$tree1,leaf.labels=c("Control","Treat"))

Code

# Tree depth 2
plot(grf_est1$tree2,leaf.labels=c("Control","Treat"))

1.4 Forestsearch with depth=2 (maxk = 2)

Code

# Setup parallel processing
library(doFuture)
library(doRNG)

registerDoFuture()
registerDoRNG()

system.time({fs <- forestsearch(df.analysis,  confounders.name = confounders.name,
                                outcome.name = "time_months", treat.name = "hormon", event.name = "status", id.name = "id",
                                potentialOutcome.name = NULL, 
                                df.test = NULL,
                                flag_harm.name = NULL,
                                hr.threshold = 1.25, hr.consistency = 1.0, pconsistency.threshold = 0.90,
                                sg_focus = "maxSG", max_subgroups_search = 30,
                                use_twostage = TRUE,
                                showten_subgroups = TRUE, details=TRUE,
                                conf_force = NULL,
                                cut_type = "default", use_grf = TRUE, plot.grf = TRUE, use_lasso = TRUE,
                                maxk = 2, fs.splits = 1000,
                                n.min = 60, d0.min = 10, d1.min = 10,
                                plot.sg = TRUE, by.risk = 6,
                                parallel_args = list(plan="callr", workers = 30, show_message = TRUE)
)
})


=== Two-Stage Consistency Evaluation Enabled ===
Stage 1 screening splits: 30 
Maximum total splits: 1000 
Batch size: 20 
================================================

GRF stage for cut selection with dmin, tau = 12 0.6 
tau, maxdepth = 46.75811 2 
   leaf.node control.mean control.size control.se depth
1          2         6.49        82.00       3.34     1
2          3        -4.10       604.00       1.06     1
11         4        -7.90       112.00       2.81     2
21         5         3.86       177.00       1.87     2
4          7        -5.89       356.00       1.33     2

Selected subgroup:
  leaf.node control.mean control.size control.se depth
1         2         6.49        82.00       3.34     1

GRF subgroup found
Terminating node at max.diff (sg.harm.id):
[1] "er <= 0"

All splits:
[1] "er <= 0"   "age <= 50" "age <= 43"
GRF cuts identified: 3 
  Cuts: er <= 0, age <= 50, age <= 43 
# of continuous/categorical characteristics 5 2 
Continuous characteristics: age size nodes pgr er 
Categorical characteristics: meno grade3 
## Prior to lasso: age size nodes pgr er 
#### Lasso selection results 
7 x 1 sparse Matrix of class "dgCMatrix"
                 s0
age     .          
meno    .          
size    0.005433435
grade3  0.178139021
nodes   0.049670523
pgr    -0.001812895
er      .          
Cox-LASSO selected: size grade3 nodes pgr 
Cox-LASSO not selected: age meno er 
### End Lasso selection 
## After lasso: size nodes pgr 
Default cuts included from Lasso: size <= mean(size) size <= median(size) size <= qlow(size) size <= qhigh(size) nodes <= mean(nodes) nodes <= median(nodes) nodes <= qlow(nodes) nodes <= qhigh(nodes) pgr <= mean(pgr) pgr <= median(pgr) pgr <= qlow(pgr) pgr <= qhigh(pgr) 
Categorical after Lasso: grade3 
Factors per GRF: er <= 0 age <= 50 age <= 43 
Initial GRF cuts included er <= 0 age <= 50 age <= 43 
Factors included per GRF (not in lasso) er <= 0 age <= 50 age <= 43 

===== CONSOLIDATED CUT EVALUATION (IMPROVED) =====
Evaluating 16 cut expressions once and caching...
Cut evaluation summary:
  Total cuts:  16 
  Valid cuts:  16 
  Errors:  0 
✓ All 16 factors validated as 0/1
===== END CONSOLIDATED CUT EVALUATION =====

# of candidate subgroup factors= 16 
 [1] "er <= 0"      "age <= 50"    "age <= 43"    "size <= 29.3" "size <= 25"  
 [6] "size <= 20"   "size <= 35"   "nodes <= 5"   "nodes <= 3"   "nodes <= 1"  
[11] "nodes <= 7"   "pgr <= 110"   "pgr <= 32.5"  "pgr <= 7"     "pgr <= 131.8"
[16] "grade3"      
Number of possible configurations (<= maxk): maxk = 2 , # combinations = 528 
Events criteria: control >= 10 , treatment >= 10 
Subgroup search completed in 0.01 minutes
Found 13 subgroup candidate(s)
# of candidate subgroups (meeting all criteria) = 13 
# of unique initial candidates: 13 
# Restricting to top stop_Kgroups = 30 
# of candidates to evaluate: 13 
Algorithm: Two-stage sequential 
  Stage 1 splits: 30 
  Screen threshold: 0.763 
  Max total splits: 1000 
  Batch size: 20 
Parallel processing: callr with 30 workers

*** Subgroup found: {er <= 0} 
% consistency criteria met= 0.93 
SG focus= maxSG 
Subgroup Consistency Minutes= 0.045 
Algorithm used: Two-stage sequential 
Candidates evaluated: 13 
Candidates passed: 7 
Subgroup found (FS) with sg_focus='maxSG'
Selected subgroup: {er <= 0} 
Minutes forestsearch overall = 0.06 
Consistency algorithm used: twostage

   user  system elapsed 
 19.018   1.572   3.740

Code

plan("sequential")


# Results for estimation (training) data, which_df = "est" is default
res_tabs <- sg_tables(fs, ndecimals = 3, which_df = "est")

res_tabs$sg10_out

Identified Subgroups
Two-factor subgroups (maxk=2)
Factor 1	Factor 2	N	Events	E₁	HR	P_cons
{er <= 0}		82	45	16	1.951	0.930
{er <= 0}	{pgr <= 32.5}	75	41	16	2.222	0.990
{er <= 0}	!{age <= 43}	68	38	14	2.164	0.960
{er <= 0}	{pgr <= 7}	64	34	13	1.992	0.920
{er <= 0}	{size <= 35}	61	34	15	2.537	1.000
{er <= 0}	{nodes <= 7}	61	31	11	2.335	1.000
{er <= 0}	!{size <= 20}	61	35	12	2.054	0.930
Search Configuration: Single-factor candidates (L) = 32; Maximum combinations evaluated = 528; Search depth (maxk) = 2
Search Results: Candidate subgroups found = 13; Maximum HR estimate = 2.54
Note: E₁ = events in treatment arm; P_cons = consistency proportion

Code

res_tabs$tab_estimates

Treatment Effect Estimates
Training data estimates
Subgroup	n	n1	events	m1	m0	RMST	HR (95% CI)
ITT	686 (100.0%)	246 (35.9%)	299 (43.6%)	66.3	50.2	7.8	0.69 (0.54, 0.89)
Questionable	82 (12.0%)	26 (31.7%)	45 (54.9%)	22.9	43.7	-14	1.95 (1.04, 3.67)
Recommend	604 (88.0%)	220 (36.4%)	254 (42.1%)	66.7	52.6	9.3	0.61 (0.47, 0.80)

1.5 Bootstrap Inference

Code

#output_dir <- "dev/vignettes-working/applications/gbsg/results"
output_dir <- "results/"
save_results <- dir.exists(output_dir)

# patchhwork needed for a combined bootstrap plot (otherwise if not avaialable will not produce)
library(patchwork)

# Number of bootstrap samples
NB <- 100

system.time({fs_bc <- forestsearch_bootstrap_dofuture(
  fs.est = fs, 
  nb_boots = NB, 
  show_three = FALSE, 
  details = TRUE)
})

Ystar matrix generated should be 'boots x N': 100 x 686

ForestSearch parameters for bootstrap iterations:
  - sg_focus: maxSG 
  - maxk: 2 
  - fs.splits: 1000 
  - max_subgroups_search: 30 
  - hr.threshold: 1.25 
  - hr.consistency: 1 
  - pconsistency.threshold: 0.9 
  - n.min: 60 
  - use_twostage: TRUE 
  - use_lasso: TRUE 
  - use_grf: TRUE 
  Bootstrap-specific overrides:
  - grf_res: NULL (forces re-selection)
  - grf_cuts: NULL (forces re-selection)
  - parallel_args: sequential (prevents nested parallelism)
  - details: FALSE (suppressed in workers)
  - plot.sg: FALSE
  - plot.grf: FALSE


=== Bootstrap Analysis Complete ===
Success rate: 87.0% (87/100)

H (Questionable) Estimates:
  Unadjusted:       1.95 (1.04,3.67) 
  Bias-corrected:  1.56 (0.71,3.43) 

Hc (Recommend) Estimates:
  Unadjusted:       0.61 (0.47,0.8) 
  Bias-corrected:  0.64 (0.5,0.81) 
===================================

   user  system elapsed 
976.284  19.248  92.981

Code

plan("sequential")


if (save_results) {
    filename <- file.path(output_dir, 
                         paste0("gbsg_k2b_B=", 
                                format(NB), 
                                ".RData"))
    save(df.analysis, fs, fs_bc, file = filename)
    cat("\nResults saved to:", filename, "\n")
}


Results saved to: results//gbsg_k2b_B=100.RData

1.5.1 Diagnostics and Summaries

Code

#load("~/Documents/GitHub/forestsearch/vignettes/results/sim_gbsg_example_B=1000.RData")

output_dir <- "results/"

#NB <- 1000

load_results <- dir.exists(output_dir)
if(load_results){
filename <- file.path(output_dir, 
                         paste0("gbsg_k2b_B=", 
                                format(NB),".RData"))

load(file = filename)
}

summaries <- summarize_bootstrap_results(
      sgharm = fs$sg.harm,
      boot_results = fs_bc,
      create_plots = TRUE,
      est.scale = "hr"
    )


===============================================================
           BOOTSTRAP ANALYSIS SUMMARY                          
===============================================================

BOOTSTRAP SUCCESS METRICS:
-------------------------------------------------------------
  Total iterations:              100
  Successful subgroup ID:        87 (87.0%)
  Failed to find subgroup:       13 (13.0%)

TIMING ANALYSIS:
-------------------------------------------------------------
Overall:
  Total bootstrap time:          1.53 minutes (0.03 hours)
  Average per iteration:         0.02 min (0.9 sec)
  Projected for 1000 boots:      15.27 minutes (0.25 hours)

Per-iteration timing:
  Mean:                          0.16 min (9.6 sec)
  Median:                        0.16 min (9.6 sec)
  Std Dev:                       0.09 minutes
  Range:                         [0.03, 0.39] minutes
  IQR:                           [0.09, 0.23] minutes

ForestSearch timing (successful iterations only):
  Iterations with FS:            100 (100.0%)
  Mean FS time:                  0.16 min (9.6 sec)
  Median FS time:                0.16 min (9.6 sec)
  Total FS time:                 15.94 minutes
  FS time % of total:            1044.2%

Overhead timing (Cox models, bias correction, etc.):
  Mean overhead:                 0.00 min (0.0 sec)
  Median overhead:               0.00 min (0.0 sec)
  Total overhead:                0.02 minutes
  Overhead % of total:           1.5%

PERFORMANCE ASSESSMENT:
-------------------------------------------------------------
  Performance rating:            ✓✓✓ Excellent
  Average iteration speed:       0.9 seconds

===============================================================

Code

sg_tab <- summaries$table

sg_tab

Treatment Effect by Subgroup
Bootstrap bias-corrected estimates (100 iterations)
Subgroup	Sample Size			Survival			Treatment Effect
Subgroup	N	N_T	Events	Med_T	Med_C	RMST_d	HR (95% CI)^†¹	HR^‡ (95% CI)²
Qstnbl	82 (12.0%)	26 (31.7%)	45 (54.9%)	22.9	43.7	-14	1.95 (1.04, 3.67)	1.56 (0.71,3.43)
Recmnd	604 (88.0%)	220 (36.4%)	254 (42.1%)	66.7	52.6	9.3	0.61 (0.47, 0.80)	0.64 (0.5,0.81)
¹ Unadjusted HR: Standard Cox regression hazard ratio with robust standard errors
² Bias-corrected HR: Bootstrap-adjusted estimate using infinitesimal jacknife method (100 iterations). Corrects for optimism in subgroup selection.
Note: Med = Median survival time (months). RMST_d = Restricted mean survival time difference. Subgroup identified in 87.0% of bootstrap samples.

Code

event_summary <- summarize_bootstrap_events(fs_bc, threshold = 12)


=== Bootstrap Event Count Summary ===
Total bootstrap iterations: 100
Event threshold: <12 events

ORIGINAL Subgroup H on BOOTSTRAP samples:
  Control arm <12 events: 0 (0.0%)
  Treatment arm <12 events: 0 (0.0%)
  Either arm <12 events: 0 (0.0%)

ORIGINAL Subgroup Hc on BOOTSTRAP samples:
  Control arm <12 events: 0 (0.0%)
  Treatment arm <12 events: 0 (0.0%)
  Either arm <12 events: 0 (0.0%)

NEW Subgroups found: 87 (87.0%)

NEW Subgroup H* on ORIGINAL data:
  Control arm <12 events: 1 (1.1% of successful)
  Treatment arm <12 events: 1 (1.1% of successful)
  Either arm <12 events: 2 (2.3% of successful)

NEW Subgroup Hc* on ORIGINAL data:
  Control arm <12 events: 0 (0.0% of successful)
  Treatment arm <12 events: 0 (0.0% of successful)
  Either arm <12 events: 0 (0.0% of successful)

Code

summaries$diagnostics_table_gt

Bootstrap Diagnostics Summary
Analysis of 100 bootstrap iterations
Category	Metric	Value
Success Rate¹	Total iterations	100
	Successful subgroup ID	87 (87.0%)
	Failed to find subgroup	13 (13.0%)
	Success rating	Good ✓✓
Subgroup H (Questionable)	Unadjusted estimate	1.95 (1.04, 3.67)
	Bias-corrected estimate	1.56 (0.71, 3.43)
	Bias correction impact²	20.2%
	CI width change³	2.64 -> 2.72
Subgroup Hc (Recommend)	Unadjusted estimate	0.61 (0.47, 0.80)
	Bias-corrected estimate	0.64 (0.50, 0.81)
	Bias correction impact²	3.4%
	CI width change³	0.33 -> 0.32
Bootstrap Quality: H	Valid iterations	87
	Mean (SD)	0.44 (0.40)
	Coefficient of variation⁴	90.0%
	Skewness⁵	0.01
Bootstrap Quality: Hc	Valid iterations	87
	Mean (SD)	-0.45 (0.21)
	Coefficient of variation⁴	45.5%
	Skewness⁵	0.27
Search Performance	Mean max HR found	3.32 (1.37)
	Mean factors evaluated	39.4
	Mean combinations tried	832
	Proportion at maxk	--
¹ Success Rate: Proportion of bootstrap samples where ForestSearch identified a valid subgroup
² Bias Correction Impact: Percentage change from unadjusted to bias-corrected estimate
³ CI Width Change: Confidence interval width before -> after bias correction
⁴ Coefficient of Variation: Standard deviation as % of mean (lower is better)
⁵ Skewness: Measure of asymmetry (0 = symmetric, \|skew\| < 1 is generally good)
Interpretation Guide: ✓ Good stability: Subgroup is reliably identified in most bootstrap samples. ⚠ High variability: Bootstrap estimates are imprecise (CV >= 25%). Consider increasing nb_boots or sample size.

Code

summaries$subgroup_summary$original_agreement

                            Metric      Value
                            <char>     <char>
1:      Total bootstrap iterations        100
2:           Successful iterations         87
3: Failed iterations (no subgroup)         13
4:       Exact match with original   5 (5.7%)
5:         Different from original 82 (94.3%)

Code

summaries$subgroup_summary$factor_presence

  Rank Factor Count   Percent
6    1    pgr    62 71.264368
1    2    age    38 43.678161
7    3   size    29 33.333333
2    4     er    18 20.689655
5    5  nodes    11 12.643678
3    6 grade3     8  9.195402
4    7   meno     1  1.149425

Code

summaries$subgroup_summary$factor_presence_specific

   Rank Base_Factor Factor_Definition Count  Percent
34    1         age       {age <= 47}     9 10.34483

Code

summaries$plots$combined

1.6 Forest Search n-fold cross-validation

Code

output_dir <- "results/"
load_results <- dir.exists(output_dir)
#NB <- 1000
if(load_results){
filename <- file.path(output_dir, 
                         paste0("gbsg_k2b_B=", 
                                format(NB),".RData"))

load(file = filename)
}

# Kfolds = n (default to n-fold cross-validations)

fs_OOB <- NULL

fs_OOB <- forestsearch_Kfold(fs.est = fs, details = TRUE, Kfolds = 30,
                             parallel_args = list(plan = "callr", workers = 36, show_message = TRUE))

Cross-validation setup:
  - Observations: 686 
  - Folds: 30 
  - Fold sizes (range): 22-23 

ForestSearch parameters for CV folds:
  - sg_focus: maxSG 
  - maxk: 2 
  - fs.splits: 1000 
  - max_subgroups_search: 30 
  - hr.threshold: 1.25 
  - hr.consistency: 1 
  - pconsistency.threshold: 0.9 
  - n.min: 60 
  - use_twostage: TRUE 
  - use_lasso: TRUE 
  - use_grf: TRUE 
  - (per-fold parallel: sequential)
  - (per-fold details: FALSE)
  - (per-fold plot.sg: FALSE)

Cross-validation complete:
  - Time: 0.28 minutes
  - Subgroup found in 60 % of folds
Any found: 0.6 
Exact match: 0.3666667 
At least 1 match: 0.3666667 
Cov 1 any: 0.3666667 
Cov 2 any: NaN 
Cov 1 and 2 any: 0 
Cov 1 exact: 0.3666667 
Cov 2 exact: NaN 
Agreement (sens, ppv) in H and Hc: 0.5121951 0.9635762 0.65625 0.9356913

Code

# Reset workers to single
plan(sequential)

summary_OOB <- forestsearch_KfoldOut(res=fs_OOB, details=TRUE, outall=TRUE)

Any found: 0.6 
Exact match: 0.3666667 
At least 1 match: 0.3666667 
Cov 1 any: 0.3666667 
Cov 2 any: NaN 
Cov 1 and 2 any: 0 
Cov 1 exact: 0.3666667 
Cov 2 exact: NaN 
Agreement (sens, ppv) in H and Hc: 0.5121951 0.9635762 0.65625 0.9356913 
        Subgroup        n              n1            m1     m0     RMST   
Overall "ITT"           "686 (100.0%)" "246 (35.9%)" "66.3" "50.2" "7.8"  
FA_0    "Not recommend" "82 (12.0%)"   "26 (31.7%)"  "22.9" "43.7" "-14"  
KfA_0   "Not recommend" "64 (9.3%)"    "19 (29.7%)"  "37.7" "42"   "-0.34"
FA_1    "Recommend"     "604 (88.0%)"  "220 (36.4%)" "66.7" "52.6" "9.3"  
KfA_1   "Recommend"     "622 (90.7%)"  "227 (36.5%)" "66.7" "52.6" "7.8"  
        Hazard ratio       
Overall "0.69 (0.54, 0.89)"
FA_0    "1.95 (1.04, 3.67)"
KfA_0   "0.91 (0.44, 1.88)"
FA_1    "0.61 (0.47, 0.80)"
KfA_1   "0.69 (0.53, 0.89)"

Code

table(summary_OOB$SGs_found[,1])


!{age <= 43}  {age <= 48}  {age <= 50}    {er <= 0} 
           4            1            2           11

Code

table(summary_OOB$SGs_found[,2])


!{age <= 43}  {age <= 47}  {age <= 48}  {pgr <= 33} 
           2            3            1            1

Code

Ksims <- 30

fs_ten <- forestsearch_tenfold(fs.est = fs, sims = Ksims, Kfolds = 10, details = TRUE, 
                       parallel_args = list(plan = "callr", workers = 36, show_message = TRUE))

Starting repeated K-fold cross-validation:
  - Simulations: 30 
  - Folds per simulation: 10 
  - Workers: 13 

ForestSearch parameters for CV folds:
  - sg_focus: maxSG 
  - maxk: 2 
  - fs.splits: 1000 
  - max_subgroups_search: 30 
  - hr.threshold: 1.25 
  - hr.consistency: 1 
  - pconsistency.threshold: 0.9 
  - n.min: 60 
  - use_twostage: TRUE 
  - use_lasso: TRUE 
  - use_grf: TRUE 
  - (per-fold parallel: sequential)
  - (per-fold details: FALSE)
  - (per-fold plot.sg: FALSE)

Repeated K-fold CV complete:
  - Time: 2.15 minutes
  - Successful simulations: 30 / 30 
  - Projected hours per 100 sims: 0.12

Code

# Reset workers to single
plan(sequential)

print(fs_ten$find_summary)

       Any      Exact At least 1       Cov1       Cov2  Cov 1 & 2 Cov1 exact 
       0.7        0.3        0.3        0.3         NA        0.0        0.3 
Cov2 exact 
        NA

Code

print(fs_ten$sens_summary)

   sens_H   sens_Hc     ppv_H    ppv_Hc 
0.3841463 0.9403974 0.4362162 0.9179549

Code

print(head(fs_ten$sens_out))

        sens_H   sens_Hc     ppv_H    ppv_Hc
[1,] 0.4390244 0.9536424 0.5625000 0.9260450
[2,] 0.4024390 0.9271523 0.4285714 0.9195402
[3,] 0.3902439 0.9354305 0.4507042 0.9186992
[4,] 0.4024390 0.9519868 0.5322581 0.9214744
[5,] 0.2682927 0.9536424 0.4400000 0.9056604
[6,] 0.2804878 0.9254967 0.3382353 0.9045307

Code

print(head(fs_ten$find_out))

     Any Exact At least 1 Cov1 Cov2 Cov 1 & 2 Cov1 exact Cov2 exact
[1,] 0.7   0.4        0.4  0.4  NaN         0        0.4        NaN
[2,] 0.7   0.3        0.3  0.3  NaN         0        0.3        NaN
[3,] 0.8   0.3        0.3  0.3  NaN         0        0.3        NaN
[4,] 0.6   0.3        0.3  0.3  NaN         0        0.3        NaN
[5,] 0.5   0.2        0.2  0.2  NaN         0        0.2        NaN
[6,] 0.7   0.1        0.1  0.1  NaN         0        0.1        NaN

Code

# Save all results

output_dir <- "results/"
save_results <- dir.exists(output_dir)

if (save_results) {
    filename <- file.path(output_dir, 
                         paste0("gbsg_k2b_CV=", 
                                format(Ksims), 
                                ".RData"))
    save(df.analysis, fs, fs_bc, fs_ten, fs_OOB, file = filename)
    cat("\nResults saved to:", filename, "\n")
}


Results saved to: results//gbsg_k2b_CV=30.RData

Code

#Ksims <- 200

output_dir <- "results/"
load_results <- dir.exists(output_dir)
if(load_results){
filename <- file.path(output_dir, 
                         paste0("gbsg_k2b_CV=", 
                                format(Ksims),".RData"))

load(file = filename)
}



#' # Define subgroups to display
subgroups <- list(
 age_gt65 = list(
 subset_expr = "age > 65",
 name = "age > 65",
     type = "reference"
   ),
 age_lt65 = list(
 subset_expr = "age <= 65",
 name = "age <= 65",
     type = "reference"
   ),
pgr_positive = list(
 subset_expr = "pgr > 0",
 name = "pgr > 0",
     type = "reference"
   ),
pgr_negative = list(
 subset_expr = "pgr <= 0",
 name = "pgr <= 0",
     type = "reference"
   )
  )


# Create the forest plot
 result <- plot_subgroup_results_forestplot(
   fs_results = list(fs.est = fs, fs_bc = fs_bc, fs_OOB = fs_OOB, fs_kfold = fs_ten),
   df_analysis = df.analysis,
   subgroup_list = subgroups,
   outcome.name = "time_months",
   event.name = "status",
   treat.name = "hormon",
   E.name = "Hormon",
   C.name = "CT",
   ci_column_spaces = 25
 )

# Display the plot
plot(result$plot)